The compact, high-powered NVIDIA MGX modular design server powered by NVIDIA GPUs and Intel Xeon processors is set to transform and advance enterprise edge AI performance. 

TAIPEI, June 4, 2024 /PRNewswire/ -- Aetina, a subsidiary of the Innodisk Group and an expert in edge AI solutions, is thrilled to announce its innovative SuperEdge NVIDIA MGX short-depth server, model AEX-2UA1, at Computex 2024. The AEX-2UA1 is one of the first x86-CPU short-depth NVIDIA MGX servers to be showcased to the market, marking a significant expansion of Aetina's product portfolio from specialized edge devices to comprehensive AI server solutions.  

The SuperEdge NVIDIA MGX server also represents a milestone in the progression of Innodisk Group's AI roadmap, extending from frontline sensors and storage devices to AI software tools, computing platforms, and now AI edge servers. The SuperEdge server delivers maximum performance within a compact form factor for edge computing and empowers enterprises to effectively handle sensitive data with private LLM (large language model) and other enterprise AI applications. It is particularly valuable in sectors such as finance and healthcare, where data privacy is paramount, and 5G telco, offering a compact, powerful and space-optimized system ideal for mission-critical applications and deployments.

NVIDIA MGX Server Accelerates On-premises Enterprise AI Deployment 

Enterprises are embracing the rapid transformations accelerated by generative AI. However, for industries that handle sensitive data, such as finance, biotech and medical research, cloud-based LLM services can pose potential operational and compliance risks. As a result, these enterprises urgently need a solution that allows LLMs to operate in a secure manner while balancing cost and performance according to their specific cases. 

In response to the increasing demand, Aetina unveils the AEX-2UA1, the enterprise-level edge server, to enable enterprises to deploy their private LLM AI training and inferencing on premises. This allows models to be trained using sensitive data and utilized securely to ensure compliance with strict regulations. 

Powered by technologies from NVIDIA and Intel, the AEX-2UA1 supports two double-deck GPUs with NVIDIA NVLink bridges and a single Intel® Xeon 6 processor, enhancing LLM training and inferencing performance with direct GPU-to-GPU interconnect to unleash powerful edge computing performance. Furthermore, the AEX-2UA1 supports advanced network solutions, including NVIDIA BlueField-3 DPUs and SuperNICs, and NVIDIA ConnectX-7 NICs. This allows efficient server-to-server and server-to-storage communication for parallel processing in enterprises edge AI systems. 

Moreover, to address the space constraints of edge environments, the AEX-2UA1 features an x86 short-depth design. With front access and the compact form factor, it optimizes space utilization, making it suitable for deployments in both rack and non-rack locations. This versatility enables the deployment of diverse AI-driven applications wherever needed at the edge. 

Additionally, the AEX-2UA1 complies with the NVIDIA MGX architecture, offering a modular server architecture for maximum flexibility and forward compatibility, including support for NVIDIA Blackwell GPUs, delivering greater scalability for enterprises in the future. 

Aetina x Innodisk: Architect Intelligence in the Edge 

James Su, Director of Software R&D at Aetina, states, "AEX-2UA1 not only marks Aetina's advancements in product portfolio extension but also serves as an ideal entry system for enterprises in need of LLM deployment, offering low barriers to entry compared to standard GPU servers in the market. We streamline the development and deployment of AI applications through this NVIDIA MGX server, paired with our extensive expertise, which saves valuable resources while accelerating time to value and empowering organizations to swiftly harness the potential of AI and gain a competitive edge." 

The launch of the new NVIDIA MGX server reaffirms Aetina's position as the driving force in AI innovation within the Innodisk Group, showcasing its strength and momentum in pushing technological boundaries. Leveraging Innodisk Group's vast expertise in cross-industry integration and insights from serving IPC clients, Aetina transforms cutting-edge innovations into practical endpoint solutions, facilitating the widespread adoption of edge AI across diverse markets.  

Additionally, Innodisk Group is enhancing its commitment to excellence by fully backing the expansion of production capacity through a forthcoming second-phase manufacturing center.

This strategic move reinforces Aetina and Innodisk's dedication to delivering top-tier AI solutions, not only to meet the demands of the evolving market but also to pave the way for future innovations.

The SuperEdge NVIDIA MGX short-depth server will be showcased at COMPUTEX 2024. Visit Aetina's booth (Booth No. J0110) to discover how this innovative server can bring business operations to the next level. 

For more information about Aetina and the SuperEdge NVIDIA MGX short-depth server, please visit https://www.aetina.com/.